Semiparametric Evaluation of Specific Surrogates of Protection for Time-to-Event Clinical Endpoints Allowing for Time-Varying Treatment Effects

• Activity Number: 470
• Type: Contributed
• Date/Time: Wednesday, August 1, 2012 : 8:30 AM to 10:20 AM
• Sponsor: Biometrics Section
• Author(s): Erin Gabriel*+ and Peter B Gilbert
• Companies: University of Washington and University of Washington/Fred Hutchinson Cancer Research Center
• Keywords: Surrogate ; Time to event ; Waning

Abstract

Specific surrogates of protection (SoP) are important targets for Phase I and II trials and can be used to estimate phase III trial outcomes prior to running an expensive large clinical trial. There is a lack of methods for SoP evaluation that allow for time-to-event clinical endpoints and to our knowledge there are none that allow for and characterize waning in vaccine efficacy or surrogate quality over time. We propose to extend the semi-parametric estimation method for specific surrogate evaluation of Huang and Gilbert Biometrics 2011 to a continuous time-to-event clinical endpoint. We use a novel parameterization of the Weibull distribution to model risk while allowing for the characterization of possible waning in the efficacy of the treatment or vaccine as well as possible waning in the surrogate quality. We extend the predictiveness curve summary statistics of Gu and Pepe IBJ 2011 for use in surrogate quality assessment. These summary measures allow for easy comparisons of candidate SoP at given times, allowing for higher quality SoP to be identified. Methods are tested via simulation and example from the Zoster vaccine trial of Merck.